Textural Analysis of Quantitative Magnetic Resonance Imaging in Metabolic Bone Disease — An Approach to Tissue Characterisation of the Spine

Previous work has demonstrated highly significant reduction of T1 and T2 relaxation time measurements of in vivo lumbar inter vertebral discs with age and disc degeneration (Jenkins et al. 1985). The results were in keeping with water loss known to occur in a disc with age, with degeneration being an extreme form of ageing. That study applied a region of interest (ROI) approach to calculate the relaxation time values from multiple data point sets using a robust non-linear least squares iterative technique (Hickey et al. 1986). When similar ROI relaxation time measurements of lumbar vertebral bodies from the same data set were analysed, these failed to show any significant discrimination due to age, sex, anatomical level or status of adjacent disc (Jenkins et al. 1989). A failure to detect changes within vertebral bodies has been ascribed to positioning errors in the ROIs, slice thickness of the magnetic resonance (MR) data set and diurnal and biological variations in heterogeneous tissue.